Parallel breadth-first search method based on shared memory system structure

Abstract

The invention discloses a parallel breadth-first search method based on a shared memory system structure. The method comprises: a worker thread grouping and binding process, a data interval partition process, and a task execution process. The method specifically comprises: obtaining a CPU topological structure of an operating environment, partitioning and binding a worker thread group; partitioning data intervals, and making the data intervals corresponding to the worker thread groups; task execution using a hierarchical ergodic method, in an execution process, automatically selecting a proper state from a series top-down state, a parallel top-down state, and a bottom-up state, as a self execution state, and when count in an activated state in a certain layer is zero, ending task execution. The method reduces communication among CPU sockets through task partition, data consistency maintenance among the CPU sockets is eliminated, and improves data reading efficiency through data binding and automatic switching, so as to improve operation efficiency of program.

Description

technical field [0001] The invention relates to a parallel breadth-first search method, in particular to a high-efficiency parallel breadth-first search task division method and task scheduling strategy on a multi-socket shared memory architecture. Background technique [0002] Breadth-first search method is an important search method based on graph structure data. Graph structure data G usually consists of a set V of vertices and a set E of edges between the vertices. Any edge e=(u, v) means that there is an edge from vertex u to vertex v. Given a root vertex r, the breadth-first search method finds all vertices reachable by the root vertex r by expanding outward through the boundary between found and unfound vertices. In this search process, the breadth-first search method builds a breadth-first search tree. The parent-child vertex relationship in the tree means that the child vertex v is reached by the expansion of the parent vertex u. The execution time overhead of t...

AI Technical Summary

Problems solved by technology

In an application where only one breadth-first search is required, this preparation time is prohibitive

In order to maintain the information of the subgraphs divided into, this method requires additional storage space

In addition, for undirected graphs, this method needs to store two copies of an edge in different forms of incoming and outgoing edges, requiring twice the memory space

For breadth-first methods, increased memory consumption limits the size of the problem that the system can solve

At the same time, this method requires additional memory to store subgraph information, and this additional space grows with the number of NUMA nodes, which makes the program's scalability limited

Images